Automated Driving Control of Vehicles with Guidances

Abstract

Automated driving control of vehicles is considered, where both longitudinal and lateral dynamics are taken into account. Firstly, the vehicle dynamics are approximated by a global linear model obtained by Koopman operator theory. Then, a linear model predictive controller is designed, in which the global linear model is used to predict vehicle dynamics. Thus, it can reduce the computational burden accordingly. The state of the preceding vehicle is used as the reference of the following vehicle to complete the automated driving control. Finally, the efficacy of the proposed strategy is confirmed through simulation.